Antonio Oliver
77 posts


@birdabo I still don't understand, even with a founder like Mark Zuckerberg, why Meta is still not launching its own AI model?
English

🚨people forget that Zuckerberg is insane at coding.
> bro started coding at 11 and built a messaging system for his dad’s dental office before AOL even did it.
> his private tutor couldn’t keep up with him.
> in high school he built an AI music recommendation engine so good that both Microsoft and AOL tried to buy it.
- he turned both down as a teenager btw. then went to Harvard, built Facebook in a few weeks, and dropped out as a sophomore to run it.
bro is now personally steering Meta’s AI strategy while most tech CEOs haven’t written a line of code in decades.
say what yall want about Zucc but the man is a beast.

English

I was tired of the bullshit mac apps for tracking deep work sessions. So I made a pretty simple python no dependency GUI that works on mac.
You can set time for each session, time for rest and tracks how much you have done daily and weekly.
It will notify you when you have completed the session.
And it stores all this info on a local sqlite database which you can query.
Pretty simple solution, and it works for me and my workflow, I leave you the repo here if its useful for anyone, pretty simple to run and no installs needed, you only need to have python installed on your computer and a Mac.
github.com/antoni0z/simpl…
Its crazy that all the simple mac apps for this are bloatware that asks for a subscription for this simple functionality.


English
Antonio Oliver retweetledi

Dev team red flags 🚩
“We do Scrum”
“We do Git flow”
“We use ”
“We have a change approval board”
“We rarely deploy”
“We don’t have a CI server”
“We don’t have tests”
“We don’t need types”
“We are in a change freeze”
“We have a separate QA team”
“We can only use Windows”
“We aren’t allowed to use AI”
“We have many long-lived environments”
“We require working in a remote VM”
“We require approval for every install”
English

@__tinygrad__ Is this compatible with multiple GPUs? If so whats the limit?
Amazing work!
English

@MaximeRivest Modular monorepo its better, you can use uv workspaces
English

@chrisarderne @gvanrossum Zod better than pydantic? Thought the opposite
Why do you say so?
English

@gvanrossum On the one hand, Python deserves better tooling: Drizzle and zod are way nicer than Python equivalents.
On the other: Python has already lost that battle to TypeScript and this PEP won't change that. Data is where python still shines. Focus on that.
English
What do y'all think of PEP 827?
PEP 827 – Type Manipulation | peps.python.org/pep-0827/
English
Antonio Oliver retweetledi


@fchollet I think the next step into that direction may be cracking continual learning.
English

By explicitly training on specific tasks, we ended up covering a very large area (in absolute terms) of the space of all possible tasks humans can do, but this large area only amounts to 0.00...01% of the total space. And that's why we still need general intelligence.
Teng Yan@tengyanAI
@fchollet what would be example of 'novel unfamiliar domains'? seems like we've touched most things
English
Antonio Oliver retweetledi

If you ever feel like you're late to the game, consider that in the 1890s many scientists thought physics as a field was completely solved (quote below is from Albert Michelson in 1894).
On the front of intelligence science, it feels more like the 1870s. For the first time we have something that is starting to really work (however primitive it may be), which we can use as a springboard for the next few decades of discoveries.

English

@fresheyeball @digitalocean Retard enterprise microservice mess enjoyers
English

I failed a "systems design" technical interview recently, because I modeled everything as a single database. They gave no reason to do otherwise.
Some people think distributed nightmare really should be the default. A single @digitalocean droplet can handle the throughput of most businesses.
Thorsten Ball@thorstenball
me, 10 years ago: plz let me work on distributed systems! now: oh god please keep it one process one machine for as long as you can
Harney, CO 🇺🇸 English

Claude remote control does not seem that ready. Gets stuck for me in errors calling tools and stays that way for the session. Dissapointing.
@AnthropicAI

English
Antonio Oliver retweetledi

Its says quite the opposite of what you are saying😂
"though evidence is observational and causation unclear"
Many people that eat a lot of eggs and meat have an unhealthy lifestyle and overconsume other stuff, vegans quite the opposite that could be one of the reasons of the "observational" results.
Observational and studies on Big populations over years are largely bullshit due to difficulty of measuring inputs and outputs with precision.
English

@F0ODHub Again notice low consumption 1 or 2 a day not a worry - but higher can be.
x.com/i/grok/share/8…
English









